|
PureBytes Links
Trading Reference Links
|
Mark:
>So I would like to ask, in all seriousness, how do low-lag moving
>averages improve the profitability of trading systems? I guess
>my confrontational stance is, "Just because they have lower lag
>doesn't automatically guarantee they are more profitable."
I can offer an example.
Mark Jurik calls JMA a "low lag" smoother, but my reverse
engineering efforts indicate that "variable lag" is a more accurate
term. Or "variable speed."
A while back, I wrote a modification to the parabolic indicator
for a successful trader (he makes his living trading). He was
using the parabolic as a trend-direction indicator to filter other
trading signals, nothing more. The problem was, if the market
trended one way, then made a sudden huge move the other way, and
resumed trending the same direction as before, the parabolic would
reverse and then take too long to settle, indicating an obviously
wrong trend direction for many bars. So I modified the parabolic to
take into account statistics of past movements, adjusting the speed
of the indicator according to a nonlinear function based on the
statistics. It worked pretty well. He's been using it in various
live-trading systems for the past 3 years or so. And I can't argue
with his trading success.
The algorithm I employed to do this is quite similar to what JMA
does. The point is, here is a situation where the use of such an
algorithm increases profits. I used the concept behind JMA to
modify the parabolic indicator. I COULD HAVE USED JMA TO ACCOMPLISH
THE SAME THING, **BUT** I would need access to JMA's internal speed
control parameter. The fact that I would have needed an open source
algorithm doesn't diminish the value of low-lag smoothers. It only
diminishes the value of black boxes.
-Alex
|